The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to dynamically discover and process content from a remote MCP server (mcp.arkiv.network), creating a surface for potential injection.
Ingestion points: The agent is guided to fetch tool definitions, prompt schemas, and knowledge text from the remote gateway using operations like tools/list, prompts/list, and read_knowledge as described in SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instruction' warnings for the content returned by the external service.
Capability inventory: The skill environment expects the agent to execute tools based on discovered schemas and provides guidance for local writes using the Arkiv SDK.
Sanitization: No specific sanitization or validation logic is provided for the dynamically fetched schemas or knowledge content before they are processed by the agent.

arkiv-mcp